AI Agents That Live Inside a Dreamed-Up World

An agent watches a game, learns to hallucinate the next frame, then plays inside its own dream — but only the model that knows players react to each other stays true.

TL;DR: The hottest idea in agents right now: don’t feed them the real world — let them dream it. An agent watches some footage, learns to hallucinate the next frame, and practices inside its own head. I built a tiny one, and found the catch: a dream only stays true if it knows the players react to each other. Runs on a laptop.

The world

Two players in an 11-cell corridor: a predator steps toward the prey, the prey steps away. Every move is a reaction. An agent watches random games, then closes its eyes and dreams 15 frames ahead, feeding each prediction back in as the next input. I built two dreamers from the exact same footage:

• single-player — predicts each player from its own position alone
• multiplayer — predicts both positions together

Training an agent inside its own learned dream goes back to Ha & Schmidhuber’s World Models (2018); the open frontier is making that dream multiplayer — modeling agents reacting to each other, not just physics.

The 10-second version

% of the dream still matching reality, this many frames ahead:

frames ahead 1 3 5 10 15 single-player dream 11 0 0 9 0 multiplayer dream 100 100 100 100 100

The single-player dream falls apart almost immediately; the multiplayer one stays locked to reality. That gap is the whole point — and, as we’ll see, it comes down to what each model is even able to represent.

How it works

real = dream = start
for _ in range(15):
    real  = real_next(*real)   # what actually happens
    dream = model(*dream)      # feed the dream its OWN last frame
    match += (dream == real)

Enter fullscreen mode Exit fullscreen mode

Same footage, same loop. The only difference: whether the model predicts the two players jointly or independently.

Why single-player collapses

The prey moves because the predator moved. A single-player model looks at each player in isolation, so it structurally can’t represent that reaction — its errors compound each frame until the dream is fiction. The multiplayer model conditions on both, so it captures the coupling. To be fair, this isn’t a surprising empirical result so much as a demonstration: the corridor is deterministic, so the outcome really follows from what each model is allowed to see. That’s exactly why the framing matters.

A dream you can act in is a superpower — an agent can practice a thousand risky moves for free. But a dream that forgets everyone else reacts to you isn’t practice. It’s a delusion.

Why it’s exciting

The proven part: training agents inside a learned dream works — from World Models (2018) to DreamerV3 (2023) mastering 150+ tasks, and Genie (2024) learning playable worlds from video alone. Dreamed worlds let agents rehearse infinitely, safely, at zero real-world cost.

Where it’s heading: those dreams are mostly single-agent today. The 2026 push is multiplayer — worlds where agents model each other. The lesson from this demo is the whole ballgame: model the reactions or the dream drifts. Get it right and agents can plan against each other entirely in imagination.

How faithful is this?

A real world model learns from raw pixels with a neural net, in a noisy, stochastic world. This is that mechanism stripped to its core — a frequency table over a tiny, deterministic game. It’s built to make the intuition concrete, not to reproduce the papers; those (linked below) do the heavy, learned version.

Try it

git clone https://github.com/Shridhar-2205/secret-lives-of-agents
cd secret-lives-of-agents/03-dreamed-world && python demo.py

Enter fullscreen mode Exit fullscreen mode

The series — The Secret Lives of AI Agents

1. Agents invent their own language
2. Agents build a culture on a decaying notepad
3. Agents that live inside dreamed-up worlds (you’re here)

Shridhar Shah — Senior Software Engineer on the AI team at Cisco. GitHub · LinkedIn

Sources & further reading: Ha & Schmidhuber, World Models (2018) — the “train inside a dream” idea · Hafner et al., Mastering Diverse Domains through World Models (DreamerV3, 2023) · Bruce et al., Genie: Generative Interactive Environments (2024).